differential-facet-verifier
Audit, compare, and verify multi-provider evidence dossiers using explicit facet availability semantics, absence claimability gates, and 3-valued differential outcomes. Do not use for generic line text diffing or single-provider analysis without differential evidence.
- License 许可
- MIT
- 最近更新
- 2026-11-15
- Stars
- ★ 0
- Forks
- 0
01 前言 Front Matter
- name
- differential-facet-verifier
- description
- Audit, compare, and verify multi-provider evidence dossiers using explicit facet availability semantics, absence claimability gates, and 3-valued differential outcomes. Do not use for generic line text diffing or single-provider analysis without differential evidence.
02 资源文件 5
- 参考文档CARD.md2.6 KB
- 参考文档config.default.yaml615 B
- 参考文档references/absence-rules.md1.7 KB
- 参考文档references/facet-schema.md2.9 KB
- 脚本scripts/facet_engine.py4.8 KB
03 SKILL.md 正文
Differential Facet Verifier & Availability Semantics
The differential-facet-verifier skill enforces rigorous epistemic comparison across multi-provider reverse-engineering and security evidence dossiers. Rather than assuming that an unobserved dimension represents zero occurrences, this skill mandates explicit facet availability maps, enforces formal absence claimability gates, decouples raw provider syntax variations from canonical semantics, and synthesizes 3-valued differential reports (unchanged, changed, unknown).
1. Trigger Bounds & Deflection
Activation Triggers
- Comparing output across multiple reverse-engineering tools (e.g. Ghidra vs IDA vs Binary Ninja vs Rizin).
- Auditing decompiler and disassembler observations for unobserved vs empty facets.
- Evaluating absence claims for symbols, vulnerabilities, or backdoor procedures.
- Verifying differential reports for strict unknown propagation.
Negative Deflection (When NOT to Use)
- Do not use for generic line text diffing — Use standard Git diff or difflib.
- Do not use for single-provider analysis without differential evidence — Use
reconstruction-obligation-ledgerinstead. - Do not use for fuzzy heuristic text matching lacking explicit provider coverage declarations.
2. Operational Progression (The Shu Stages)
| Stage | Name | Key Activity | Completion Gate |
|---|---|---|---|
| 1 | Declare & Ingest | Ingest observation dossiers; parse facet_availability and coverage. |
Every facet carries boolean available; available: false has non-empty reason. |
| 2 | Validate Absence | Evaluate formal absence claimability rule against coverage state. | Claims rejected unless status == 'complete' and clean traversal. |
| 3 | Project Dimensions | Project comparable structural facets into canonical sorted representations. | Syntax variations recorded in limitations; representations normalized. |
| 4 | Evaluate Outcomes | Map dimensions to 3-valued statuses (unchanged, changed, unknown). |
Zero unresolved facets assigned non-null comparison counts. |
| 5 | Synthesize & Seal | Aggregate overall status and compute canonical SHA-256 summary digest. | DifferentialReport emitted with reproducible rep:<sha256> digest. |
3. Diagnostic Coaching Scorecards
| Dimension | Pass Condition | Fail Symptom | Repair Action |
|---|---|---|---|
| Facet Conflation Guard | Empty facet has explicit { available: true } |
Missing/empty facet treated as 0 occurrences | Force status: unknown, set counts to null, append explanation to limitations |
| Absence Claimability Guard | absence_claimable == True before claiming absence |
Absence asserted when truncated == True or status != 'complete' |
Reject assertion; return unresolved with limitation reason |
| Mandatory Reason Invariant | Non-empty string for every available: false |
reason: null or whitespace on unavailable facet |
Hard-block via __post_init__; raise ValueError at construction time |
| Format Decoupling Guard | Raw syntax deltas isolated from structural semantics | Provider disassembly syntax differences flagged as functional change | Record providersDiffer limitation notice; equate only normalized canonical projections |
| Digest Determinism Guard | Identical inputs produce identical rep:<sha256> |
Non-canonical JSON drift across runs | Enforce canonicalize_json for all digest inputs |
4. CLI Execution Seam
Execute via the unified harness CLI:
# Compare two multi-provider observation dossiers
harness facet compare left_ghidra.json right_rizin.json --provider-differ --format rich
# Validate formal absence claimability
harness facet claim-absence coverage_dossier.json
# Audit differential report integrity
harness facet verify differential_report.json --strict
# Export machine-parseable JSON certificate
harness facet export differential_report.json
Or execute via the co-located runner script:
python scripts/facet_engine.py compare left.json right.json
Anti-Patterns
- Facet Conflation — Treating an unobserved or undeclared facet as an empty collection (
[]), producing false zero-deltas in dimension comparison. - Undeclared Absence Fallacy — Asserting that an entity is absent when the provider's traversal coverage was truncated, partial, or
unknown. - Provider Text False Contradiction — Equating raw provider formatting (e.g. differing disassembly syntaxes) and asserting functional disparity rather than projecting structural semantics.
- Limitation Erasure — Silently dropping failure and limitation reasons when propagating unresolved dimension outcomes to the summary report.
- Premature
unchangedVerdict — Concludingunchangedon the overall report when any dimension carriesunknownstatus — overall verdict must be the strictest across all dimension outcomes.
5. References & Documentation
04 同仓库更多 Skills
ui-behavior-primitive-architect
Architect, audit, and verify zero-dependency headless UI behavior primitives, WebMCP agent bridges, WAI-ARIA/SSR/RTL tri-modal checks, and framework expert routing. Do not use for static CSS theme styling or backend database migrations.
trustgraph-ontology-governor
Govern knowledge graphs, SHACL-AF triple policies, OntologyRAG subclass validation, PROV-O dual named graphs, and Laya entropy decision scoring. Do not use for relational SQL queries or generic web scraping.
shim-harness-governor
Govern universal CLI agent loops, 4-tier tool permissions, ancestor-symlink workspace sandboxes, deterministic-first command routing, collision-safe MCP namespacing, and layered skill loading. Do not use for static UI styling or external SQL analytics.
graphiti-zep-memory-governor
Govern bi-temporal knowledge graphs with Graphiti future-edge clamping, MinHash/LSH entity dedup, RRF/MMR reranking, and Zep typed ontologies. Do not use for static SQL joins or generic web crawling.
agora-mcp-workbench-architect
Architect, audit, and orchestrate Agora Workbench MCP servers, loopback sidecars, unified state graphs, gateway/dispatcher connectors, and policy-enforced data lakes. Do not use for static frontend CSS styling or non-MCP desktop GUI automation.
skill-eval-assurance-runner
Evaluate and assure agent skills using small-N Wilson confidence intervals, unbiased pass^k reliability, deterministic trajectory graders, preflight boundary honesty, and sealed digest verification. Do not use for subjective prose review or unconstrained production deployment.